【作者】 张倩； 郑丽； 卢洋； 李林； 郭秀艳；

【Author】 Zhang Qian;Zheng Li;Lu Yang;Li Lin;Guo Xiuyan;School of Psychology and Cognitive Science, East China Normal University;

【机构】 华东师范大学心理与认知科学学院；

【摘要】 识别和学习他人的动作对个体的社会交互具有重要意义,内隐学习对这一能力的习得起关键作用。远距离规则是一种复杂的不相邻规则(如映射规则),可以探讨排除样例、组块等表面特征的影响下内隐抽象规则的习得过程。本研究以光点运动序列表达的正立与倒立的人体动作为材料,探讨人体动作远距离规则是否内隐习得。实验材料包括八个人体动作,使用该八个动作材料构成长度为10个动作为一组动画序列,以人体朝向的左右对称规则为内隐规则,即在每组动画序列中,前五个人体方向与后五个镜像对称(即映射规则)。实验程序分学习和测验两部分。在学习阶段,研究者给两组被试分别随机呈现48组符合对称规则的正立或倒立(由正立的人体动作镜像垂直翻转而来)的人体动作的动画序列,重复呈现3遍。在测试阶段,要求被试在看完一组动画序列后,首先判断刚刚看到的动画序列是否符合规则,随后从"猜测、直觉、记忆和规则"四个选项中选择他们本次判断的依据,猜测和直觉为无意识结构知识的指标,记忆和规则为意识结构知识的指标。结果显示,对于总分类正确率,正立的人体动作分类正确率高于随机水平(0.5)与倒立的人体动作,而倒立的人体动作分类正确率与随机水平差异不显著。对于意识性测量,正立和倒立的人体动作条件下意识结构知识的选择比例均小于1%,表明被试主要是根据无意识的结构知识进行分类判断;对于选择为无意识结构知识的选项,正立的人体动作分类正确率高于随机水平、倒立的人体动作,而倒立的人体动作分类正确率与随机水平差异不显著。这些结果表明,与倒立的人体动作相比,被试较容易内隐地习得正立的人体动作的左右对称规则。个体能够内隐地习得人体动作的对称规则,为内隐学习中的抽象性问题提供了新的证据。更多还原

【Abstract】 Effective processing and learning human actions of others can facilitate our social interaction. Moreover, implicit learning is of great importance to process such movements. The nonadjacent regularities are complex and about nonlocal rather than adjacent elements(e.g., inversions), which could help to explain what sortof structure can be implicitly learnt with controlling both chunks and repetition structures. In the present study, we examined whether symmetry(inversion relation) of upright or inverted human actions could be learned implicitly. We adopted point light displays, which isolate human kinematics by depicting human activity through a simple set of light points(e.g., 13 points). Participants were presented with ten actions where the directions of the first five actions predict those of the last five by mirror symmetry(i.e., inversions), e.g., the first action orients to the right, then the sixth action orients to the left, and so on. In the training phase, participants in two groups were respectively presented with 48 grammatical upright or inverted action sequences, presented three times in random order. During test phase, they were required to judge each sequence whether the given sequence was grammatical and attribute their decision basis to four categories(guess, intuition, memory and rules). Guess and intuition attributions were combined as indicators of unconscious structural knowledge, and memory and rule attributions were combined as indicators of conscious structural knowledge. For the classification performance, the proportion of correct response for upright actions was higher than chance(0.5) but not inverted actions. Moreover, the proportion of correct response for upright actions was higher than inverted actions. For the consciousness measurement, response proportions of conscious structural knowledge in the upright and inverted action conditions were less than 1%, suggesting that learning was more implicit. For the unconscious structural knowledge, the proportion of correct response for upright actions was higher than chance and inverted actions. Taken together, these findings suggest that people can implicitly learn symmetry rule of upright actions but not inverted actions, presenting a challenge to existing models of implicit learning.更多还原